Agenda

"Machine Learning-based Pattern Recognition, e.g. for Document Analysis"

Stadelmann Thilo
ECLT Fellow
Professor at ZHAW School of Engineering and Scientific Director at ZHAW digital

03 July 2020, 3 p.m.
Go here to participate through Zoom Meeting!

Abstract
Following ECLT's recommendation to get to know the members, this seminar serves two purposes: (a) to quickly introduce my group at the ZHAW Zurich Unversity of Applied Science in Switzerland, which focuses on dee learning-based pattern recognition for diverse use cases; (b) to spice this up, the largest part of the presentation will revolve around two concrete examples from our portfolio of industrial use case-driven research. We will present the challenges of industrial-strengh optical music recognition (OMR) as well as print media monitoring, and our novel deep learning-based solutions. The goal of this presentation is to introduce you to a novel object detection method that excels at finding many tiny objects in large images (also relevant in e.g. medical image processing), and to break the ice for potential future exchange and collaboration.
 

Bio sketch
Thilo Stadelmann is professor of computer science at the ZHAW School of Engineering in Winterthur, Switzerland, and the scientific director of ZHAW digital. He received his Ph.D. from Marburg University, Germany, in 2010 for his research in multimedia analysis and voice recognition, and worked in different engineering and leadership roles within the automotive industry for several years. His current research focuses on applications of deep learning to diverse kinds of pattern recognition tasks on audio and visual data from the medical and industrial domain, where robustness and interpretability are key. Thilo is the co-founder and former head of the ZHAW Data Science Laboratory, board member and co-founder of the Swiss Alliance for Data-Intensive Services and a Fellow of the ECLT European Centre for Living Technology in Venice, Italy.